Typically, estimating an angle at which a character is drawn is performed using an Optical Character Recognition (OCR) process. This process is slow and not suitable for real-time OCR applications. Existing solutions either ignore such angle or request the user to specify the angle manually, which can be burdensome for the user especially with a high volume of text that needs to be processed.